Speaker driver measurements: 2-4" midrange, full range

[TMM]

Hi all,

Several years ago I measured a bunch of 2-4" cone midrange / 'full-range' drivers and published the results on my website in a 'Zaph Audio'-style. You can view the measurements and commentary here:

http://feleppa.com.au/speakermeasmid.html

I still have all the measurement equipment I used in 2013 and I'm hoping to get back into measuring soon. I plan to add +10dB and -10dB harmonic distortion sweeps in addition to the existing test level otherwise the test methodology won't change.

My interest is still primarily in small drivers with the focus on using them as midranges in passive 3-way systems. Drivers I have acquired and plan to test very soon are as follows:
-Scanspeak D7608/920010 3" dome midrange
-Dayton Audio RS52AN-8 2" dome midrange
-Dayton Audio RS52FN-8 2" dome midrange
-Dayton Audio RS75-4 2" full rangeÂ*
-Aura NS3-193-8A1 3" full range
-Peerless TC5FB00-04 1.5" full range
-Peerless TC5FB00-08 1.5" full range
-Peerless TC5FC00-04 1.5" full range
-Peerless TC5FC07-04 1.5" full range
-Peerless TA6FC00-04 2" full range
-Peerless TA6FD00-04 2" full range
-Peerless TC7FD04-04 2.5" full range
-Peerless 830970 2" full range
-Tang Band W2-748SG 2" full range
-DLS Ultimate U2.5 and UR2.5 2" dome midranges (Morel CDM54 and a version with a sealed rear chamber)
-Tang Band W3-2141 3" full range
-SB Acoustics SB65WBAC25-4 2.5" full range
-HiVi B2S 2" full range

All of the drivers previously tested and the ones listed above have been purchased by myself with funds from my own pocket. I have no affiliation with any audio companies, my day job is not in the audio industry. Up until this point I have been doing this only to satisfy my own curiosity. If there is significant interest in a particular driver and I become interested too, i'll consider purchasing and testing it. Being in Australia shipping is quite killer from USA and Europe so I can often only justify buying a whole bunch of drivers at once, rather than buying just one pair at a time.
I may consider in the future accepting driver donations for testing if there is demand. I will never do a 'paid review' however - my opinion can't be bought and I will always call a spade a spade

I may also consider doing ~1" tweeters and ~6.5" woofers in the future as I have a bit of a collection of those now too .

Any feedback is always appreciated. Cheers!
-Tim

[Bear]

Fun stuff! The RS100 often falls into this camp, as does the NE123. I've always been curious about the Morel mid-dome, but I haven't seen it in too many designs. It's been around a long time, though, so it must be doing something right!

I'm always grateful for someone spending their money to keep me informed!!

[TMM]

The ones I have are the car audio Morel dome mid. I'm not sure if they are different to the MDM55 (home audio morel dome) but I'd wager that they probably have the same or very similar internals.

New driver measurements added:
-Dayton Audio RS52AN-8 2" dome midrange
-Dayton Audio RS52FN-8 2" dome midrange
-Aura NS3-193-8A1 3" full range
-Peerless TC5FB00-04 1.5" full range
-Peerless TC5FB00-08 1.5" full range
-Peerless TC5FC00-04 1.5" full range
-Peerless TA6FD00-04 2" full range
-Peerless TC7FD04-04 2.5" full range
-Peerless 830970 2" full range
-Tang Band W3-2141 3" full range

Link to results

All the peerless measured as expected. The TC5 are kind of interesting; Tweeter-like response at the top end but much different distortion characteristic at the bottom end since they are an overhung motor with much higher xmax and lower fs than the average tweeter. They overload more gradually than the average tweeter. Would work great in an array - I'm thinking of how I can use a pair of them with a 7" woofer.

Peerless 830970 and Aura NS3 are close to unlistenable imo.

The Tang band has an excellent motor, I just wish they found a better cone for it. I've always said that TB need to put a metal cone on one of their better motor designs. This motor with a metal cone would be killer.

The RS52F is a bit of an upset. The RS52A I purchased years ago performs much better.

[PANDINUS]

Great work !
I appreciate sharing the information with other fans of the speakers. Thank you very much.

[Bear]

The RS52F is a bit of an upset. The RS52A I purchased years ago performs much better.

The price one pays to get rid of the ~15K breakup node, I guess. Presumably a crossover would be cheaper for the fabric dome than what's required to tame the aluminum dome, but at $40, there are a lot of competing alternatives for that duty (e.g., ZA14W08), and the total cost of using the 10F/8414 is probably less in most designs as well, despite a $24 price gap.

And ... Thanks!

[TMM]

Zaph has a neat solution for dealing with the RS52A breakup in his ZDT3.5 design; just shunt the existing lowpass inductor with a small cap to create a notch filter. The disadvantage of doing this versus making a dedicated RLC notch filter is that the lowpass filter ceases to continue rolling off above the notch frequency, in fact the transfer function even starts rising back up. It's not so much an issue in this case because the response is already pretty far down at that point and the RS52 frequency response is steeply rolling off too.

http://www.zaphaudio.com/ZDT3.5.html

More results hot off the test baffle:
-Scanspeak D7608/920010 3" dome midrange
-Dayton Audio RS75-4 2.5" full range
-Tang Band W2-748SG 2" full range
-HiVi B2S 2" full range

Also added HD measurements at higher levels for the Fountek FR88EX, enjoy!

Link to results

[TMM]

Driver measurements added:
-DLS Ultimate U2.5 (Morel CDM54) 2" Dome
-SB Acoustics SB65WBAC25-4 2.5"
-Dynavox LY302F 3.5"
-MCM Audio Select 55-5650 3.5"
-Peerless TA6FC00-04 2"

Link to results

Really impressed with the Dynavox LY302F - probably the best 'full range' driver I've tested to date. The MCM 55-5650 would also make a fine midrange - and I suspect it is a close relative of the Dayton RS100

[Bear]

Good stuff!

[5th element]

It's interesting that my measurements of the SB65WBAC at 92dB do not show the increase in distortion that yours do.

Now mine show a distortion rise sooner in frequency than yours, presumably because, in my case, these were taken in a tiny box. Maybe the increased air pressure of the small box changes the way the air interacts with the voice coil former vents. Or it might not be from the vents at all. Something is definitely going on though.



I also limited my measurements to 300hz as the cones were flapping about so much I didn't want to risk damaging the driver.

[TMM]

Yours doesn't seem too far off mine.

The main increase I'm seeing between 86dB and 92dB levels is that the 3rd harmonic between 400 and 800Hz increased from about -55dB to about -45db but at 300Hz it stayed put. Your 92dB measurement seems to be showing about the same general level of 3rd order distortion over that frequency range. The main reason I pointed it out was that the behaviour is not really in line with a motor simply being pushed towards it's excursion limit. When you get a bit of a 'hump' in the distortion curve like that which isn't caused by cone breakup nodes, it typically points to a suspension or aerodynamic problem. Problems like this are bad because the distortion level increases very quickly with increasing volume and there are often significant high order distortion components introduced. It seems to be more common on smaller drivers for some reason - Tang Band W2-748SG does it, Peerless PLS-P830970 does it, Peerless TA6 does it, even the Fountek FR88EX does it. Air trapped behind the dustcap/cone is my guess.

By contrast, something like the MCM 55-5650 is a good example of a driver that doesn't suffer any of these issues - instead we just see all order of distortion smoothly increasing with decreasing frequency and increasing level. Peerless TC5 is another which seems to get away without much of this kind of issue. Edit: Dayton rs75 also.

Btw your plot will have higher distortion around 300Hz because of an FFT windowing limitation near the start of the sweep. I have the same issue on mine - the plotted distortion right at the start of every sweep is higher than actual. You can see the effect on a simulated distorted sweep at the bottom of this page:

http://feleppa.com.au/speakermeas.html

The main reason I wrote my own scripts was because HolmImpulse doesn't do a very good job at windowing near the start of sweeps. ARTA seems to be doing ok though.

[5th element]

I just went and re-examined my data. I took distortion measurements at 2.83VRMS, 8VRMS and 0.83VRMS. So similar in essence to what you've done.

Like you said the increase from 0.83 > 2.83VRMS doesn't really give much of an increase in the higher order products, but going up to 8VRMS does. With my measurements though I limited things to 300Hz @8VRMS because the cone was flapping about so much I figured I was going to be exceeding the devices xmax and I didn't want to risk damaging anything.

The TC5 looks to suffer the same increase in higher order harmonics to me, in fact it looks worse overall than the SB65. The RS75 does look considerably better and needs pushing another 6dB for its harmonics to become noticeably higher. This could indeed be because of the phaseplug.

As I mentioned above the excursion of my SB65 was quite high, especially down at 300Hz. It's possible that the issue arises from when the voice coil moves backwards. When it does the perforations, in the voice coil former, start to become blocked off by the felt pad and the pole piece, certainly not a good situation to be in. I guess the only way to really see what's what would be to remove the dustcap and replace the felt pad and cap with a phase plug. The only issue with this is that the dust-cap probably contributes a reasonable amount to the drivers SD and with a phase plug in place the perforations in the former would have an easy path into which they could chuff between themselves and the plug.

One thing you can see in my measurements is that the higher order HD products (4th and 5th) remain flat down to 500Hz, below which they rise steadily. This is perfectly normal behaviour. Yours does seem to have some nasty peaks at around 500Hz.

Of course our measurement conditions are completely different, my tiny box, and presumably your free air ones. It would make sense for my distortion to show the usual rise earlier than yours do.